Ministry of Water
Resources, Works
and Housing
WORKSHOP
To Design the Technical Investigation of the Potential to Reoptimize the
Akosombo and Kpong Dams to Restore Ecological Functions and
Livelihoods in the Lower Volta River Basin
AT THE KOFI ANNAN CENTER, ACCRA, MAY 29 & 30, 2007
Key Highlights and Conclusions from Workshop
SESSION 1:
DEFINING A TARGET DOWNSTREAM FLOW PATTERN TO RESTORE ECOLOGICAL FUNCTIONS AND
LIVELIHOODS

Involvement of the downstream communities is essential to assure that the river is restored to meet
their preferences and needs. The traditional leaders—Chiefs and Elders—can play a key role in
organizing the community involvement, together with academic experts in the “human ecology” of the
Lower Volta River.

One element of this process should be reconstructing the realities of lives and livelihoods before
Akosombo dam was closed in 1966. Thus, there is a special role to be played in this process by the
elders in these communities.

However, new lifestyles, livelihoods, floodplain land uses and settlements have emerged since the
flow has been regularized. Some community members, particularly those who never experienced a
dynamic river, may prefer some elements of current life along the river.

Thus, it is probably unrealistic to seek to attain a consensus vision for the river among the community
members. Rather, the project should seek to define a range of degrees of restoration of the river,
develop reoperation scenarios to achieve these flow patterns and leave the ultimate choice of futures
to the exogenous decisional processes.
1

A range of flow patterns should be defined in any event to reflect the substantial year-to-year
variations in rainfall and the flows that occurred under natural conditions. Perhaps dry-, medium- and
wet-year flow targets should be specified.

There are two standard methodologies that are commonly applied to define environmental flow
regimes. A combination of these should be applied in the Lower Volta context:
1. Bottom up: biohydrologic and human ecology objectives are specified. In this case, restoration of
the shell fishery, the fin fishery, floodplain agriculture, and geomorphic processes to shape the
channel, connect the river to the estuary, flush sediments from the channel and restore beach
morphology, groundwater replenishment, etc. The best empirical data is then marshaled to relate
these objectives to flow patterns and hydrodynamic conditions, or, more realistically, ranges
thereof.
2. Top down: the pre-development hydrograph is specified from river gauge and reservoir inflow
data, covering the range of hydrologic conditions reflected in the period of record. Perturbations
expected from climate change should also be factored in. This hydrograph is then broken down
into key components. Components such as flood flows that could cause catastrophic damage in
the floodplain or impinge upon settlements or land uses to an unacceptable degree are then
eliminated from the hydrograph. The remaining flow pattern—varied to reflect annual differences
in hydrologic conditions, is then used as the target for reoperation flows. The underlying
assumption is that flows as near as possible to the natural pattern will support the broadest array
of ecological processes and environmental services.
o Key analytical tools that will be needed for this process are digital elevation models that
can relate degrees of floodplain inundation to releases at the dam(s) and
hydraulic/sediment process models that can indicate the magnitude of flow releases
needed to mobilize and transport sediments below the dam(s).
SESSION 2:
CONSTRUCTING AND EVALUATING VARIOUS OPERATIONAL SCENARIOS TO
ACHIEVE THE TARGET FLOWS

VRA stated that it would prefer to have the project develop a new planning model evaluating
operational scenarios than to use the VRA’s current dam operating models.

The project will need a suite of linked models and a sequence of such models. It will be most efficient
to evaluate scenarios at a coarse level initially for screening purposes, and then pursue more detailed
and refined analysis of the most promising and robust scenarios using more sophisticated models.

The types of models that will be needed include:
o A basin mass-balance water resources management model to allow “what if” gaming of
scenarios;
o An operational optimization model that will include all of the current operational objectives plus
achievement of the environmental flow releases;
o A flood routing (hydrodynamic) model that will show how various magnitudes of releases
interact with the floodplain;
2
o Rainfall-runoff (hydrologic) models to estimate the inflows into the system (because the river
gage data are probably not adequate for this purpose);
o A hydraulic sediment process model;
o Perhaps a surface water/groundwater interaction model;
o Hydrobiological process model(s) (if there are sufficient data to construct such a model).

The workshop debated whether a basin-wide planning model would be necessary to account for the
storage and diversion of water upstream from Lake Volta, especially in light of the various plans for
future development of water resources (e.g., Bui dam). It was expressed that this would be desirable
and might be accomplished without extensive effort by adapting and augmenting the models that are
being developed by other research projects on the Volta basin.

It became apparent that the process of selecting models for use in this project would require a more
intensive exercise by the modeling experts and that this should constitute an explicit and early task in
the scope of work for the project.

A limiting factor analysis will be needed for a number of potential constraints on reservoir reoperation,
including:
o An assessment of the engineering and economic feasibility of retrofitting the dam(s) with
additional outlet and turbine capacities;
o An assessment of the floodplain constraints on the magnitude of flows that can be
reintroduced. The DEM and flood routing modeling will indicate what inundation patterns
impinge upon particular land uses and structures. The extent to which these constraints can
be modified to accommodate flood releases will be an exercise in judgment that may call for
developing alternative scenarios and sensitivity analyses;
o An assessment of sediment availability constraints in recreating geomorphic processes.
SESSION 3:
HOW TO CONSTRUCT A MODEL OF THE CURRENT POWER GENERATION AND DISTRIBUTION SYSTEM
AND EVALUATE SCENARIOS FOR CHANGING THE SCHEDULING OF POWER OUTPUT FROM
AKOSOMBO AND KPONG

The key insight from the workshop was that Akosombo/Kpong represent such a large fraction of the
power generation in the current grid system that the feasibility of reoperation may depend critically
upon its integration into a more regionalized grid, as is currently being pursued under the auspices of
ECOWAS.

Corollary to this insight is the conclusion that the project should adapt and utilize the power demand
and supply optimization model that has been developed for the West African Power Pool unit of
ECOWAS by the Energy Center at Purdue University, working closely with the Ministry of Energy of
Ghana and the Volta River Authority. (The project could use such an economic optimization model to
explore the economically non-optimal scenarios associated with achieving environmental restoration
flows. Stated another way, the ECOWAS optimization model would need to be broadened to include
additional costs and benefits associated with this environmental performance objective.

Organizing this element of the project will require additional outreach and interaction with these
institutions, and is an urgent priority.
3
SESSION 4:
HOW TO EVALUATE THE REOPERATION SCENARIOS FOR ECONOMIC FEASIBILITY

Project would use valuation methods that account for all relevant costs and benefits, including both
those quantifiable in economic terms and those environmental, social and aesthetic effects that are
not.

The project will include in the cost-benefit analysis the following factors, among many others:
o Effects on power supply reliability. Here, the project will establish as a baseline the current
degree and characteristics of power reliability. We note in this regard that the degree of
reliability is relatively low due to the dominance of hydropower facilities that depend upon an
inherently and widely variable hydrology.
o The costs associated with a robust program of monitoring the effects of experimental
environmental flow releases against expectations (working biohydrologic hypotheses),
evaluating these effects, and determining further operational refinements.
o Compensation for changes in land use, flood proofing of structures, and resettlement of
floodplain dwellers.
o Intangible “quality of life” benefits associated with attracting back to the downstream
communities the migrants who have left over the past several decades.
o Health benefits (and detriments) must be included in the economic analysis, even though
some of these may be difficult to quantify monetarily.
o Risks and uncertainties associated with new flow regimes will need to be acknowledged
explicitly.
o The extent to which reoperation scenarios either mitigate or exacerbate the (power, water
supply and ecological) risks associated with global climate change also must be dealt with
explicitly and quantified to the extent possible.

The increase in the costs of power generation and transmission associated with rescheduling power
outputs from the dams will likely be capital costs (for additional turbines, thermal generators and
transmission capacities) rather than operating costs (for fuel). These can be represented in current
values by reference to the cost of capital (interest rates). However, the cost-benefit ratio may depend
critically upon assumptions as to whether these rates are commercial or concessionary. That will
depend upon the type of lending institution assumed.

Distributional impacts of costs and benefits will also be taken into account. For instance, the benefits
of improvements in fluvial geomorphology are likely to accrue to the beaches of Togo and Benin as
well as Ghana. It may also prove to be the case that the price of improved livelihoods for the
4
downstream residents will be paid in part in the form of (slightly?) higher power rates for other
customers.

Revenues to VRA (and through it to the public fisc) are not likely to be affected as current power rates
do not vary according to the time value of generation (i.e., peak and baseline generation produce the
same revenues). Thus, rescheduling of power is likely to be revenue neutral.
5